Sample frameDay 214·Growth · Retain

Good morning. Engagement isn't one number. It's a shape — and most companies have never looked at theirs.

Growth.

The Power User Curve

5 min read·Apply by lunch

The question

If you plotted every one of your users by how many days out of the last twenty-eight they actually turned up, would there be a bump at the far right — a genuine core who use this most days — or one long slide from the left with nothing at the end? Most businesses find the slide the first time they look. It's a much cheaper thing to discover this morning than in a board meeting.

The idea

In August 2018 Li Jin and Andrew Chen, both then investing at Andreessen Horowitz, published a piece called The Power User Curve, arguing that one chart tells you more about engagement than any summary number will. Building it takes an afternoon. Take a month and, for every user, count how many days in it they were active. Then plot the histogram: days active along the bottom, share of users up the side. Jin and Chen drew theirs over thirty days; plenty of teams use twenty-eight instead, because four whole weeks removes the weekday-weekend distortion. What comes out is a shape, and the shape is the finding. Consider two products. In the first, every user drifts in about one day a week. In the second, a fifth of users are there almost daily and everyone else appears once. Those two can post near-identical monthly engagement ratios and they aren't remotely the same business: one has a habit at the top, the other has a long tail of tourists. Jin and Chen's name for the healthy version is a smile — heavy on the left, where new and casual users always sit, but turning up again at the far right. Most curves have no right end at all.

The framework

01
Build it before you argue about anything else.
This is a spreadsheet job, not a data project. Export every user who did anything in the last twenty-eight days with a count of distinct active days, bucket them into bands of equal width, and chart the bands. Equal width matters more than it sounds — a wider band at the right-hand end manufactures the exact bump you're hunting for. Do it twice, once for people who joined this quarter and once for people who've been with you a year, because those two curves usually describe different products. And define an active day honestly: opened the app is attendance, not activity. Count only the days something landed.
02
Three shapes, three different businesses.
A curve that slides down from the left and hits the floor early is a product people try and abandon: the mass sits in the first band or two, and nothing follows. A smile — heavy on the left but turning up again at the far right — means a real core exists alongside the churn, and that core is what you build on. A curve weighted right, most users active most days, is a utility, and it's rare. Read yours as a diagnosis rather than a score: a fat left tail is an onboarding problem, a missing right end is a habit problem, and the two need opposite work.
03
The right-hand end is a description of your product.
The users who turn up twenty-five days out of twenty-eight are the closest thing you have to a definition of what this is for. Pull a few dozen of them and find what they share: the job they're doing, the trigger that brings them back, the segment they sit in, the one feature they live inside. Very often it's a use case nobody planned, and it beats any amount of survey work, because it's revealed behaviour rather than stated preference. Then ask the harder question — what would have to change for more of this month's signups to look like these people by their second month?
04
Is a daily habit even the right target here?
Not every product should be chasing the right-hand end. A payroll tool used twice a month, a tax product used once a year, a booking site used when somebody travels — each has a natural frequency set by the underlying need, and grading it against a daily curve produces engagement features nobody asked for. Decide the natural cadence first, then read the histogram against that: for a genuinely weekly product, four days in twenty-eight is a power user. A curve weighted left isn't a failure on its own. The failure is a curve weighted further left than the need itself would explain.

Members only · 4 principles + template + AI mentor

364 more frameworks are waiting.

The Power User Curve is Day 1 of 365. One framework every morning for a year — across strategy, sales, negotiation, leadership, and more.

$1 /day

Billed $365/yr · cancel any time

365 frameworks

12 topics · templates · AI mentor

Reserve your place — $365/yr →

One payment a year · the whole library · no auto-renewal.

More Growth frameworks

All 24 growth frameworks →